The influence of perception on platoon splitsby Craig Glaser
June 21, 2010
After reading Max Marchi's fantastic article about platoon splits, I thought that it would be an interesting subject to examine through the use of batting eye. Unfortunately I don't have Max's pitch classifications set up so I'm going to stick to examining the four most frequently used pitch types: four-seam fastballs, curveballs, change-ups and sliders.
We all know that platoon splits exist, but what exactly causes them? This issue has been looked at before, but I think that there is still work to be done, as there are a number of different explanations floating around. For example it could be that the break of a pitch influences the quality of contact which a batter makes, or it could be that ability to perceive the pitch. The batter might have more time to pick up a pitch when they have a platoon advantage, or the break of the pitch could make it easier to see.
If pitch perception is the key then we should be able to find platoon splits by looking at swing behavior. In order to test this I'm going to use batting eye. In addition to batting eye I'm going to include the raw components of batting eye (swing percentages at balls and strikes) and selectivity to give us more data to look at. As usual I'll be looking only at pitches on two strike counts, ensuring that batters were properly motivated to swing.
I’m going to start with fastballs. It would be logical for a platoon advantage to show up in fastballs if reaction time is the issue, since batters must react to them most quickly. Having the platoon advantage could give the batter a little extra time to pick up the pitch and decide whether to swing or not.
Looking at the numbers of right-handed batters you might be tempted to say that they demonstrate a platoon advantage when facing fastballs. They swing at more strikes and less balls against lefties than they do against righties, giving them a boost in batting eye. Left-handed batters, however, show a similar split. It seems that the biggest factor in batting eye is not the platoon match-up but simply the handedness of the pitcher.
Right-handed pitchers have an advantage when throwing fastballs, which is somewhat surprising. Maybe this is due to the existence of "crafty lefties." Perhaps the average left-handed pitcher just has a slower fastball than the average right-handed pitcher. It will be interesting to dig deeper into this question, controlling for fastball velocity, and see if that explains the difference found here.
Perhaps it is not the speed, but the break of the pitch which causes a platoon advantage in perception. Here is the swing behavior which batters exhibit on curveballs.
There are some differences here, with batters swinging a bit more at strikes when the pitcher is of the opposite handedness. These batters also swing more at balls, however, which leads to almost equivalent batting eye scores (and lower selectivity scores.) Left-handed pitchers seem to have a slight overall advantage when it comes to curveballs, though it is not as big as the advantage which right-handed pitchers have with their fastballs. There is more of a platoon split in selectivity than in batting eye when it comes to curveballs, with batters actually being more selective when facing pitchers of the same handedness.
I've shown before that batters tend to be more productive when adopting a more selective approach. The combination of this reverse platoon split in selectivity for curveballs and Max's finding of a reverse platoon split on curveball run value seem to complement each other.
Sliders show a very interesting pattern, with left-handed batters showing a rather large traditional platoon split and right handed-batters showing a smaller reverse platoon split. It, therefore, seems as though left-handed pitchers have an advantage in sliders similar to the one which right-handed pitchers have when throwing fastballs. Unlike in curveballs this doesn't match-up very well with Max's results.
Finally we’ll take a look at change-ups. Pitchers rarely throw change-ups against batters of the same handedness, especially left-handed pitchers. Does batting eye provide evidence that this is a wise strategic decision?
Right-handed batters show an almost equal batting eye score whether the pitcher is right or left-handed. The components of the score are quite different, however, with right-handed batters swinging more often at change-ups from left-handed pitchers both inside and outside of the strike zone. Left-handed batters, on the other hand, have a huge platoon split, swinging at more pitches in the zone, and less pitches outside of the zone against left-handed pitchers. This gives them a rather huge platoon split when it comes to change-ups. This does seem to validate the strategy of left-handed pitchers rarely throwing change-ups to left-handed hitters.
Unfortunately all of this adds up to a rather inconclusive story. I'd like to do some further analysis, perhaps borrowing Max's pitch classifications (especially once he perfects classification for left-handed pitchers) and see if anything more conclusive shows up. I've also started looking at individual players, which can add another level to the analysis. One example of this can be seen in my article at Amazin' Avenue "What's Wrong with Wright" where I show that David Wright had an equal batting eye score against righties and lefties in 2008 but had a much worse batting eye against right-handed pitchers in 2009 while improving against left-handed pitchers.
At this point it seems as though right-handed pitchers have an advantage when throwing fastballs while left-handed pitchers have an advantage on breaking pitches. It doesn't seem like pitch perception could be the main cause of platoon splits but it does seem like it could be contributing, especially in curveballs.
My instinct is that some combination of contact quality and pitch perception are at play here. This might explain why the results for curveballs (with mostly vertical movement) match up rather well with Max's run value results while my results for sliders (with mostly horizontal movement) do not. Accounting for velocity and amount of movement will hopefully clear up the story and help reveal the cause of platoon splits.
Read more of Craig's work on his website Sabometrics